Human colon cancer cells.

Human colon cancer cells.

Urbain Weyemi, Christophe E. Redon, William M. Bonner

Cancer immunotherapy takes aim at mutation-riddled tumors

New immune system-boosting cancer drugs have in clinical trials saved the lives of many people with seemingly untreatable melanoma or lung cancer, but the drugs seem useless against colon cancer. One exception—a man with colon cancer whose metastatic tumors vanished for several years after he was treated in 2007—piqued researchers' interest. They suspected his recovery might have to do with the large number of mutations in his tumors. Now, a small clinical trial suggests that even cancer patients with types of tumors that were thought to be impervious to the new drugs could benefit if those malignancies have the right error-riddled DNA signature, a result that could help 3% to 4% of cancer patients.

The drug tested is an antibody that blocks a receptor called PD-1 on the surface of the immune system’s T cells. Tumor cells can hide from T cells by activating the PD-1 receptor; when this immune “checkpoint” is blocked by a PD-1 inhibitor, however, the T cells see the tumor cells and can attack them. Such drugs are among two new types of immune system–harnessing cancer treatments that have generated tremendous excitement, because in some patients with advanced cancer, they keep tumors at bay for years. One hypothesis about why melanoma and lung cancer respond best to PD-1 inhibitors is that these two cancer types tend to have more mutations than other cancers. Some of these mutations may alter genes so that they code for small stretches of abnormal proteins that the immune system sees as foreign proteins, or antigens. The more mutations, the more of these “neoantigens” that can trigger an attack from T cells that have been unleashed by a PD-1 inhibitor.

When researchers at Johns Hopkins University in Baltimore, Maryland, examined tumor tissue from the original man with colon cancer who responded to a PD-1 inhibitor, they found a clue: His tumor had mutations in “mismatch repair” genes, so-called because their encoded proteins fix errors in DNA bases when cells replicate their DNA. When these genes don’t work properly, they can lead to cancer-promoting mutations and result in a colon tumor riddled with 1000 or more mutations, 10 to 100 times the usual number. The Hopkins group wondered if patients with many cancer types who had tumors with errors in mismatch repair genes would respond to PD-1 inhibitors.

To explore this idea, Hopkins oncologists Dung Le, Luis Diaz, and others looked for mismatch repair mutations in tumor samples from patients with advanced colon cancer and other cancer types whose tumors had stopped responding to other treatments. They divided 48 patients into two groups, those with and without these mutations in their tumors. Then they gave all of them a PD-1 inhibitor called pembrolizumab (Keytruda) every 2 weeks.

The difference in the results was dramatic. Those with mismatch repair mutations were far more likely to respond—of 13 with colorectal cancer, eight saw their tumors shrink, four remained stable, and only one got worse. By contrast, not a single one of 25 colorectal patients lacking mutations in mismatch repair genes responded. Some of the patients who responded are still alive after a year or more, whereas the nonresponders lived on average only 7.6 months. Of 10 patients with other types of cancer—including pancreatic, prostate, and uterine—who had mismatch repair mutations, in seven their tumors shrank or remained stable; the other three progressed. The study will appear tomorrow in The New England Journal of Medicine, and Le will present updated results at the annual meeting of the American Society of Clinical Oncology in Chicago tomorrow.

The results suggest that the 3% to 4% of all cancer patients with mismatch repair defects in their tumors could benefit from PD-1 inhibitors, Diaz says. “It’s a slice. It’s not all cancer,” he says. Still, that could add up to 30,000 to 40,000 advanced cancer patients a year in the United States, he says.

The study also supports the idea that tumors with more mutations—whether from mismatch repair defects or for other reasons—are more likely to respond to PD-1 inhibitors and similar drugs, says immunotherapy cancer researcher Jedd Wolchok of the Memorial Sloan Kettering Cancer Center in New York City. His team has recently reported that melanoma and lung cancer patients with more neoantigen-coding tumor mutations are more likely to respond to immune checkpoint blockers. The Hopkins study “gives us very good confidence that what we’ve seen before is important,” he says. One implication is cancer patients with few mutations in their tumors might respond better to PD-1 inhibitors if they first receive radiation or chemotherapy, because these treatments can create new mutations in tumors. Although some patients in PD-1 inhibitor trials have had such prior treatments, without a trial designed to address the question, “it’s hard to draw conclusions,” Diaz says.